Welcome to our NDPR review forum on Thomas Mulligan’s Justice and the Meritocratic State (Routledge 2018), recently reviewed in NDPR by Peter Dietsch. Please feel free to comment on any aspect of the book, the review, or the discussion below!
From the book blurb:
Like American politics, the academic debate over justice is polarized, with almost all theories of justice falling within one of two traditions: egalitarianism and libertarianism. This book provides an alternative to the partisan standoff by focusing not on equality or liberty, but on the idea that we should give people the things that they deserve.
Mulligan sets forth a theory of economic justice—meritocracy—which rests upon a desert principle and is distinctive from existing work in two ways. First, meritocracy is grounded in empirical research on how human beings think, intuitively, about justice. Research in social psychology and experimental economics reveals that people simply don’t think that social goods should be distributed equally, nor do they dismiss the idea of social justice. Across ideological and cultural lines, people believe that rewards should reflect merit. Second, the book discusses hot-button political issues and makes concrete policy recommendations. These issues include anti-meritocratic bias against women and racial minorities and the United States’ widening economic inequality. Justice and the Meritocratic State offers a new theory of justice and provides solutions to our most vexing social and economic problems. It will be of keen interest to philosophers, economists, and political theorists.
Excerpt from the NDPR review by Peter Dietsch:
At the core of Thomas Mulligan’s theory of justice lies the idea that one deserves social advantages — jobs and income in particular — on the basis of one’s merits. These merits depend on context, and thus the perfectionism of Mulligan’s theory is formal rather than substantive: “meritocracy remains agnostic about what is good and instead establishes a framework under which the good — no matter what it be — can best be pursued.” (p.37) The only obvious constraints that Mulligan puts on this framework are the aboutness and fitness principles: I only deserve something if the desert basis is about me in the relevant sense, for instance if I wrote this essay rather than copying it from a friend; and the social advantage I receive must fit, or be proportional to, the achievement or contribution in question. In addition, the notion of desert defended is pre-institutional and, as most but not all conceptions of desert, backward-looking.
At the level of distributive principle, Mulligan translates this core intuition into two steps. First, fair equality of opportunity is a necessary condition for a meritocracy; for example, when Daisy gets into Yale because of her privileged background rather than because of her academic excellence, then fair equality of opportunity is violated, and Daisy’s desert-claim to that Yale spot is undermined. Second, against the backdrop of fair equality of opportunity, we should judge people strictly on their merits. The project of the book is to spell out what this means in different contexts. The final two chapters of the book discuss some of the policy implications of this meritocratic ideal.
How does desert relate to other values? Mulligan declares that he is a pluralist about the morality of economic life — justice, need, and efficiency all represent relevant ideals, even though he regards justice as primus inter pares — but a monist about justice, which is grounded in desert alone. He also suggests that justice as desert happens to be particularly apt at promoting efficient outcomes.
Throughout the book and in chapter 3 in particular, Mulligan highlights the intuitive appeal of his approach. He cites studies showing that meritocratic ideas are widely accepted across the political spectrum. Therefore, Mulligan not only sees the potential of overcoming the traditional theoretical divide between egalitarians and libertarians, but of actually obtaining public support for meritocratic reforms.
The book has many virtues. First and foremost, it makes a welcome contribution to a still underdeveloped field of research. With some notable exceptions (e.g. Olsaretti 2004; Miller 2001; Sher 1987), desert-based theories of justice have not been getting the positive attention they deserve (!). Second, it is refreshing to see a book that does not shy away from attempting to bridge the gap between theory and political practice, with all the dangers and contingencies that this endeavor entails. Third, whether or not one is ultimately convinced by its arguments, the book is written in a clear and engaging style. Mulligan’s position on the many issues he addresses is never in doubt.
Having said all this, the bulk of this review article will critically discuss some arguments where Mulligan’s reasoning appears perhaps less convincing or where more development is needed. Doing so will also provide the reader with a flavor of the details of Mulligan’s case for meritocracy.
The most fundamental worry about the proposed theory is that Mulligan’s theory of merit remains underdetermined. We can see why this is so by distinguishing two conceptions of meritocracy. The first, thin conception of meritocracy is the one advertised in the book. It amounts to the basic idea that everyone should get their due. This idea is relatively uncontroversial, which strengthens Mulligan’s claim that it is indeed robustly anchored in people’s intuitions about justice (see chapter 3). The second, thick conception of meritocracy is one that provides the substantive criteria of desert necessary to ascertain people’s desert claims in particular contexts. Here, people regularly disagree. Just how much money does a CEO or a music star deserve to make? Who is the most deserving candidate for that job? As can be gleaned from the quote in the opening paragraph of this review, Mulligan eschews such a conception. However, without it, the practical import of his theory arguably remains limited.
Think of an analogy. All theories of justice agree on the importance of equality. However, their agreement is limited to a thin conception of equality as equal consideration. It does not extend to distributive questions. While utilitarians, libertarians, and egalitarians all endorse equal consideration, they disagree on its implications for the distribution of social advantages. I fear that Mulligan might face a similar challenge. Scaling up the meritocratic norm from relatively simple contexts such as the ultimatum game to complex social issues forces one to take a stance on issues that invite controversy.
I thank Sukaina for arranging this discussion, and Peter for his thoughtful review of my book. There is much we might discuss. For now let me limit my comments to two issues Peter raises: (1) underdetermination and (2) Rawls’s famous rejection of desert.
A theory of justice may be underdetermined in two ways. It might as an ideal matter fail to answer questions that we are confident fall under the ambit of justice. And this is indeed a problem with the theory. But it might also, as a non-ideal matter, fail to give plain guidance for how we should change public policy. Now, unlike some theorists, I take these non-ideal worries seriously, devoting the third and final part of the book to public policy and an early section (2.4) to discussing ideal/non-ideal conceptual issues.
Peter seems unsatisfied with these efforts. I think the necessary resources are there. One purpose of the firefighter example (5.1) is to show that giving people what they deserve consists in equal opportunity AND assessment based on merit. That is the ideal. And a lesson for our non-ideal circumstances is that we must pursue the two in tandem.
A world in which people are assessed strictly on their merits and these merits result entirely from family wealth is an unjust world. A world of perfect equal opportunity and non-meritocratic discrimination (discrimination on the basis of race, gender, appearance, etc.) is similarly unjust. We should aspire to a world in which (1) no one is advantaged or disadvantaged by what her ancestors did or did not do, AND (2) we judge people strictly on their merits. That is justice. We realize that ideal in our actual world by BOTH pursuing equal opportunity (through, e.g., confiscatory estate/inheritances taxes and massive public investment in children) and reducing discriminatory policies, like affirmative action. (AA is especially objectionable because, as a blunt instrument, it provides advantage to some people who already enjoy undeserved advantage–viz. blacks born into wealth.)
Relatedly, Peter says that I do not “aim at a positive theory of contribution, but merely a negative one”, but this is not correct. I show (6.3) that the notion of paying people in accordance with their economic contributions is well-defined; that challenges to this distributive rule which have in the past been raised are surmountable (by, e.g., using Shapley value), and that the system is Pareto efficient. What I caution the reader against is making the inference that if we wish to pursue justice (as described) then we must know precisely what each person contributed. That inference is bad. There are many non-ideal steps available to us that would doubtless make our society more just from the meritocratic point-of-view. (E.g. higher top marginal income tax rates to ward off rent extraction and Pigovian taxes).
Peter asks “what level of inequality is compatible with desert?”, as if a failure to answer this question is a flaw in the theory. In fact we should be looking at things in the opposite way: Inequality is a symptom of injustice rather than an injustice in itself (people’s contributions are not as unequal as our current income distribution). If you want to know what just inequality looks like, you find out by giving people equal opportunity and paying them in accordance with their contributions.
I turn to say a few words about Rawls. It is true that his theory permits “income premia based on natural assets”. (I might put it slightly differently: Rawls permits income premia on the basis of traits–intelligence, diligence, etc.–that are partly genetically-determined.) As Peter rightly points out, this is because it is theoretically possible and almost surely true in practice that paying such premia is in the interest of the worst-off, making more output available for redistribution.
But I think my disagreement with Rawls is not trivial. Note first that our justifications are different: Rawls thinks such premia are instrumentally useful; I think they are deontologically necessary. If you make an contribution to the economy, and that contribution emanates from natural traits, that does not diminish your desert.
Second, Rawls’s position legitimates morally dangerous remuneration. Consider the problem of taste discrimination (Becker 1957): If a society is racist, say, then output may be increased, and the material position of the worst-off improved, if we pay wage premia to members of the dominant race. The Difference Principle would have us pay such premia. And it would countenance hiring less-qualified members of the dominant race for similar reasons. The meritocrat disagrees: Race is irrelevant from the point-of-view of merit, and therefore it must be ignored in hiring, even if that makes the worst-off worse off than they might be. The meritocrat believes that that is a price worth paying.
Thanks to Sukaina for providing this forum to continue our discussion on Tom’s Justice and the Meritocratic State.
Tom’s defense in reaction to my criticism of underdetermination is that underdetermination can occur in either an ideal-theory context or a non-ideal theory context, that underdetermination in the former is more worrying, and that he takes my charge of underdetermination to be directed at what his position in non-ideal contexts.
Which of the two kinds of underdetermination is more worrying, I would think, depends on the declared goal of the theory in question. For a theory such as Tom’s that “take[s] these non-ideal worries seriously” (see 2nd paragraph of Tom’s post), why would underdetermination in non-ideal circumstances be less problematic?
Be that as it may, let’s look at the two examples of possible underdetermination in turn. First, the firefighter example: It is true that my review suggested I consider the underdetermination to show up in non-ideal contexts (“…I find this position puzzling at least in the non-ideal circumstances where equality of opportunity is violated.”). Yet, I believe that the underdetermination is already present at the level of ideal theory. “We should aspire to a world in which (1) no one is advantaged or disadvantaged by what her ancestors did or did not do AND (2) we judge people strictly in their merits.” (paragraph 4 of Tom’s post) Point 1, that is, equality of opportunity, is uncontroversial. The problem lies in point 2. What it means to judge people by their merits is underdetermined. Another way to frame the criticism is in terms of circularity: Any attempt to define what meritocracy is by appeal to the same concept is doomed to fail. What we need is precise criteria of what constitutes merit.
Unsurprisingly, this underdetermination shows up at the level of non-ideal theory: I do not see how Tom gets from (1) and (2) above to the conclusion that affirmative action, for example in the firefighter example, is unjustified. In his post, he rightly states that affirmative action will lead to some unintended consequences – e.g. providing advantage to blacks born into wealth. But this drawback of the policy surely has to be weighed against the good the policy does in terms of helping genuinely disadvantaged individuals. Again, I am not saying I know the answer to the empirical question of how these two effects balance out, but I do not see how Tom’s theory get him to his negative conclusion on affirmative action. The reason for this gap, I suggest, lies in the underdetermination of the theory.
Let’s turn to the second example: income inequality. First of all, I agree with Tom that “[i]nequality is a symptom of injustice rather than an injustice in itself” (paragraph 6 of Tom’s post). From this perspective, it would be a mistake to adopt an outcome-oriented theory of distributive justice. Now, suppose that “challenges to the distributive rule which have in the past been raised are surmountable” (paragraph 5 of Tom’s post) – I am not as optimistic as Tom that concepts such as Shapley values can in fact accomplish this feat, but let’s assume so for arguments sake. In any case, and in contrast to the firefighter example, the Shapley values do indeed provide us with an account of what constitutes merit.
Yet, this leaves two important questions unanswered. First, how do we arbitrate conflicts between different substantive accounts of what merit in terms of productive contribution consists in? In other words, how do we deal with pluralism about justice? Second, and less importantly, in light of Tom’s insistence that people’s intuitions about justice converge with his meritocratic theory, do we have any evidence about people’s intuitions on what Shapley values require?
Finally, two reactions on what Tom has to say on Rawls: First, what income inequalities, if any, can be justified by differentials in partly genetically-determined traits? I don’t see where Mulligan and Rawls come apart on this question. That said, I accept Tom’s point that they do differ in the justification of these inequalities: For Tom, one’s productive contribution is tightly linked to one’s identity and, therefore, one has an entitlement to the relative value of this contribution. For Rawls, because natural assets are arbitrary from a moral point of view, there is no such entitlement and distributive justice instead is governed by other considerations, namely the difference principle.
Second, I disagree that “Rawls’s position legitimates morally dangerous remuneration” (Tom’s last paragraph). Wage premia on racist grounds would be ruled out by the first principle of justice – equal liberty – in Rawls’s account, which has priority over the difference principle.
Peter,
Thanks again for taking the time to discuss these matters. I am especially glad that you raise this possibility that race-based wage premia (e.g.) might be ruled out by one of the lexically prior Rawlsian principles. This is something I’ve been thinking about of late; there may be a interesting paper to be written on how the Difference Principle (DP) irreparably conflicts with Equal Liberties (EL) or Fair Equality of Opportunity (FEO). (I know that Andrew Mason (2006) has written such a paper, though I have not read it yet.)
Now immediately one will say: Who cares? EL and FEO have lexical priority over DP, and so there is no problem in the event of conflict between them. However, there IS a problem if DP becomes vacuous; if interpreting EL/FEO to avoid distasteful results like race-based wage premia leaves nothing for DP to do.
So, along these lines: It seems to me, first, that it cannot be EL that rules out these premia. That principle applies primarily to the political, not the economic, sphere, and its requirements are enumerated by Rawls: freedom of speech and assembly, the rule of law, integrity of the person, etc.
On the other hand, one might naturally say that Rawlsian FEO rules out race-based wage premia. One immediate complication is that Rawls is not clear about what, exactly, FEO means to him (“rather cryptic”, as Lindblom 2018 puts it). When he says that “those with the same level of talent and ability and the same willingness to use those gifts should have the same prospects of success regardless of their social class of origin”, that sure sounds like a traditional, “level the playing field” view of equal opportunity and not a rule governing a just wage. But you need such a rule if you want to disallow taste discrimination-type remuneration.
But let’s grant for the sake of argument that Rawlsian FEO does proscribe such remuneration. What about income differences due to gender? Presumably proscribed too. The “beauty premium” (Hamermesh and Biddle)? Unclear. How about preferential hiring/wage for a family member or a friend? Aren’t we butting up against EL if we try to prevent that?
The Rawlsian can go one of two ways here. Either (1) make FEO mean Formal EO (i.e. purely meritocratic discrimination) as well, or (2) permit some forms of non-merit-based remuneration which the meritocrat would not. If (1), then Rawls’s theory collapses into meritocracy and DP is vacuous. If (2), the question is whether we are willing to live with that discrimination in order to benefit the worst-off. I am not.
I admire Rawls above all other contemporary philosophers, but lest this just become more Rawlsiana, let me make one other point. To wit, I think you overestimate the appeal of Fair Equality of Opportunity and underestimate what a victory it would be, philosophically and politically, to reach consensus about its virtue.
First, all right-libertarians reject FEO, and left-libertarians at best endorse a tepid form of it. Second, utilitarians do not require FEO as a conceptual matter (although given its efficiency benefits, they probably would endorse it in practice). Third and most important, FEO has disappeared–at least here in the U.S.–from egalitarian thought and rhetoric. It has been supplanted by calls for diversity, equal representation, and equality of outcome. Many policies that the egalitarian left endorses, such as reparations for slavery, badly violate FEO. As you know, I view this as a dangerous development, and one not likely to redound to the benefit of the American lower and middle classes.
Tom,
Thanks to you for writing the stimulating book that gives us occasion to have these discussions!
If we delve into Rawls exegesis, it is indeed an interesting question where in his system race-based wage premia would be dealt with. I am not sure I agree with you that the first principle could not play this role. There are two issues here: First, perhaps less interestingly, did Rawls himself have in mind a principle strictly limited to political liberties, and thus excluding an issue such as this one? Second, is there a principled reason for not extending the remit of the first principle to the economic sphere within a broadly Rawlsian framework? I don’t see one.
The difference principle would not become vacuous if one made that move, because it would continue to adjudicate the wage differences stemming from differences in talent.
I do think that Rawls is often misunderstood with regards to his position on the role of talents. Dworkin, for instance, thought that the difference principle did not go far enough, because it did not complete the job of rendering the distribution of social advantage “endowment-insensitive.”
The point I was trying to make in my review is that on the issue of talents, it seems to me that your beef is much more with liberal egalitarianism 2.0 à la Dworkin than with Rawls. Many if not most liberal egalitarians today agree with Dworkin that differentials in talent constitute an undeserved dis-/advantage and should thus have no impact whatsoever on the distribution of social advantage. But Rawls deliberately did not go as far as this. His position, well summarized in the sentence you cite in the book (“The natural distribution is neither just nor unjust; nor is it unjust that persons are born into society at some particular position. These are simply natural facts. What is just and unjust is the way that institutions deal with these facts.” [Rawls 1999: 87]), focuses on the way our institutions react to the distribution of natural talents. This is something he has in common with your meritocratic account, even if differences do of course remain. Perhaps surprisingly in light of what Rawls says about desert, he might be an ally of a meritocrat when it comes to the link between talents and wages.
So much for today…
Perhaps it is worth saying a few words about the concept of merit, and whether it is inevitably underdetermined. I do not believe, pace your remarks, that we need “precise criteria of what constitutes merit” to make the idea of merit-based hiring (say) coherent, or to operationalize the theory. After all, we do not have precise criteria of what constitutes beauty, but it seems reasonable to think that there is such a thing as beauty. And beauty pageants seem to do a decent job of identifying beautiful people. But I do agree that we need criteria useful enough to constrain hiring decisions, and to make non-trivial changes to economic practice. Let me give three examples of how meritocracy does this.
Example one. Consider a case of taste discrimination. A black applicant, B, and a white applicant, W, apply for a job at a profit-maximizing widget factory. B can produce 100 widgets/hr. W can produce 80 widgets/hr. But because the firm’s workers are racist, hiring B would induce a productivity loss in them, to the tune of 30 widgets/hr. So the firm hires W instead, thus maximizing profits.
Note two things. First, as a conceptual matter, it does not seem right to say that B is less meritorious than W. What might be said, rather, is that in this case the firm has reason to hire a less meritorious applicant. There is a nugget of merit that exists independent of the consequentialist goal of profit. Second, meritocracy requires that B, not W, be hired—and this is not a trivial consequence from the economic point-of-view.
Example two. Roger is the sole owner of a restaurant. Roger is a white supremacist, and the purpose of his business, which he has made explicit, is not just to make money but to hire, pay, and advance members of the white race. A black man applies for a job at the restaurant, and he is better than all other applicants when it comes to remembering orders, being attentive to customers, not dropping food, etc.
Here too it appears, to me at least, that the black applicant more meritorious; that is, merit is independent of the goals of those who own the relevant position. In other words, the concept of merit is robust to libertarian ownership.
Example three. Michael, a man, and Wanda, a woman, apply for an academic job in a field in which women are underrepresented and sexual harassment is rampant. All the committee members agree that Michael is more intelligent, that he has produced more and better-quality research, and that he is a more dedicated and skillful teacher. But in order to effect positive change in the field, they hire Wanda.
Again, whatever the virtues of hiring Wanda, they are not related to Wanda’s merit. They are related to her instrumental value. And it is my intuition that Michael remains the more meritorious applicant, in the same way that, in the first example, the black applicant is more meritorious even though he’s not the profit-maximizing one.
This is evidenced by our intuitive ability to respond to merit, and to recognize when justice is done and when it is not. Think of how Wanda would feel, after being hired, in the face of two different justifications for her selection: (1) “Michael is a better researcher, but we hired you because you are a better teacher. And for us, teaching skill is a more important trait (and it will have the following good effects . . .)” (2) “Michael is a better researcher, but we hired you because you are a woman. And we believe that gender is a more important trait (and it will have the following good effects . . . )”
To my mind, these two justifications will be received very differently, will affect our applicant, her sense of dignity, and her ability to take satisfaction in the scarce good which she has won in two different ways. This is because if (2), she (and Michael too, of course) were treated unjustly. If (1), justice was done.
As I argue in the book, I believe that we can reach consensus on many traits—including race, gender, appearance, and sexual tastes—that are irrelevant from the point-of-view of merit in virtually all contexts. The meritocrat demands that we ignore these traits when hiring. That constrains this important part of economic practice, and moves us closer—although I admit not exactly to– the ideal, which is to judge each person only on her merits.
These three examples are helpful. I agree with pretty much everything in your remarks on them. Yet, here is why I nonetheless think that more is needed for a full-fledged meritocratic theory.
Such a theory has to deliver two things. First, an account of what counts as meritorious and what does not. We have a reasonably good sense of what does not count, or at least what should not count: race, gender, appearance, and sexual tastes, for example. People should have equal opportunities irrespective of their traits in these categories.
Second, we need not only a consensus on the traits that do count towards merit, but we also need to agree on the relative weights of the different dimensions of merit. It is in this second context that things are often more complex than producing 100 widgets versus 80 widgets. People disagree about which dimensions of merit should receive more weight, and they also differ in their assessment of merit in different candidates. As an illustration, think of the last academic hiring committee you sat on. It seems to me that a meritocratic theory, in order to be a serious contender to serve as a comprehensive theory of justice, has to say something to meet these challenges.
Your book helpfully reminds us that these questions have been neglected in the literature in recent years. It also makes some important steps in the right direction. I look forward to your future work to take us further down this path.
Hi Tom (if I may),
I haven’t read the book, so my apologies if you address this there. But I have a question about your idea that the ideal society would realize two principles (equal opportunity + merit-based assessment), and that “in non-ideal circumstances… we must pursue the two in tandem.” Is this “tandem” requirement supposed to rule out reforms that violate the merit-based assessment principle in the service of helping to realize equality of opportunity (or vice versa)? For example, suppose we live in a world in which equality of opportunity is violated (as we do), and in which affirmative action policies help promote equality of opportunity (as many think, but perhaps you don’t). Is your view intended to rule out the possibility that such a reform is justified from the outset, by its requirement that we pursue each principle in tandem? Or does the justification of the reform turn on considerations such as (a) the extent to which the AA policy promotes equal opportunity, (b) the extent to which it violates the merit-based assessment principle, and (c) the relative weight of each principle as applied to this case?
A related case is one where violating the merit-based assessment principle now facilitates the eventual achievement a society whose institutions or members better conform to it in the future. (Compare Rawls’s discussion of non-ideal cases where we must restrict basic liberties in order to pave the way for the full realization of the liberty principle at a later date.) I suspect that many would place AA policies in this category, but I am more concerned with the general point than the (largely empirical) details of the AA example. On the one hand, if you say that reforms cannot violate the merit-based assessment principle either (i) to help realize equality of opportunity, or (ii) to help more fully realize the merit-based assessment principle in the future, then I worry that the position straightjackets efforts to promote reform (though perhaps you think this worry is overstated). On the other, if you think that it is okay to violate the merit-based assessment principle in such cases, then I do not see how you can directly apply the principle to policies in our decidedly non-ideal world. In other words, I am not sure what prevents someone from granting you your ideal, but then adding that, given the current state of our society, we need all sorts of transitional policies that violate merit-based assessment and so would be never realized in the ideal (though, again, I have not read the book, so maybe you are fine with this result).
Best,
Jacob
Jacob–
Thank you for the comments. I do discuss this issue in the third part of the book. As I see it, there is no conceptual hurdle to sometimes allowing discrimination (e.g. AA) in order to equalize opportunities (or vice versa). One can imagine such a circumstance: a policy which is a teensy bit discriminatory but at the same time enormously improves matters from the point-of-view of equal opportunity.
Of course, we want to know what to do in the real world. So let me say a few things about that. First, I do not find compelling evidence that modern day discrimination serves such a noble purpose. Even when they are explicitly intended to have an equal opportunity effect, discriminatory policies end up being not just ineffective but counterproductive. One example is preferential university admissions in order to provide “role models”. This has led to the “mismatch effect”: Many of those granted admission on the basis of race, rather than merit, wash out. This, in turn, perpetuates nasty (and false) racial stereotypes, undermines the confidence of young members of the community, and foments resentment in the labor market.
Second, these programs sap the public’s will for more serious, more radical, and more effective reforms. There are direct and proven methods for equalizing opportunity, like estate/inheritance taxes and investment in public education. These should be exhausted before we consider discrimination. The reason we have such disparate economic outcomes between the races is mostly (not entirely) a result of the fact that the mean black child grows up in a household with 1/10 the wealth of the mean white child. Unequal outcomes are a result of unequal opportunity—not (primarily) a result of racism, nor a result of economically-relevant genetic differences. We really should be focusing on the opportunity piece in my judgment.
Furthermore, policies like AA require virtually no sacrifice on the part of the economically dominant class. In contrast, the aforementioned equal opportunity policies require considerable sacrifice. In this way, AA provides cover for those who would rather avoid the important and radical economic reforms that justice demands. AA is unjust both in itself and in the change it precludes.
Another problem with AA as it is almost always practiced is that it provides aid on the basis of race (or gender, etc.)—not economic disadvantage. Of course, the two may be correlated. But if these policies are really meant to have an equal opportunity effect, there is no reason why we should rely on proxies for disadvantage. Doing so leads to injustice: Some people who are born into advantage gain further advantage because of the color of their skin, and some people born into poverty are saddled with more burdens (because they belong to a race which is, on average, relatively advantaged).
So in sum, while I think that the point you raise is an apt one, I do not see an urgent need for “transitional policies that would violate merit-based assessment”. Insofar as these policies already exist, they are ineffective in promoting equal opportunity. And there are many policies that directly promote equal opportunity, yet to be tried or implemented to their fullest degree, that we should use first.